COSA:NETs guidelines/Histopathology

From Cancer Guidelines Wiki
Cosa prev button.png
Cosa overview button.png
Cosa next button.png

NETs banner new.png


Gastroenteropancreatic neuroendocrine tumours (GEP NETs) arise from a variety of gastrointestinal neuroendocrine system (GI NES) cells and may behave differently according to the cell of origin and site of the tumours.

In general, GEP NET cells show a uniform morphology with ‘salt and pepper’ nuclei, low mitotic rate and uniform size and shape. However poorly differentiated tumours with less readily recognisable features are encountered.

The diagnosis of GEP NETs should be made using routine haematoxylin and eosin-stained sections in which the typical morphology of NETs is evident. The diagnosis should be confirmed by immunohistochemistry.

Back to top

Additional stains

While some NETs are readily recognisable, others are less so and some, particularly those with acinar growth patterns, may be mistaken for adenocarcinoma.

A panel of immunoperoxidase stains should be applied. As a minimum, chromogranin A and synaptophysin should be performed in all suspected GEP NETs. This may be adequate for the majority of NETs but this panel may occasionally need to be expanded to include CD56 (N-CAM), CD57 (Leu7), PGP9.5, as not all NETs express all markers. The NET panel should also be used for suspected combined glandular or squamous/endocrine tumours.

In the case of a metastatic NET, additional immunoperoxidase studies may help to determine the site of the primary tumour. The panel could include TTF-1, CDX-2, PDX-1 and NESP-55 to distinguish between pulmonary, pancreatic, foregut and hindgut NETs.[1] CK7 (foregut tumours, stomach, duodenum, pancreas) and CK20 (midgut and hindgut tumours) may be added to the panel.[2]

Although TTF1 is expressed by well-differentiated pulmonary tumours, TTF1 may however, also be expressed in 44%-80% of non-pulmonary small cell carcinomas and cannot therefore be relied on for identifying a poorly differentiated NET of pulmonary origin.

Immunostains for specific hormones may be helpful in certain circumstances but are not mandatory.

Back to top

Role of proliferative indices and mitotic rate

Ki-67 and mitotic counting should be reported as standard as these have been shown to have prognostic significance and are part of a new proposed grading[3][4] (see below). The Ki-67 stain correlates well with size, angioinvasion and behaviour of the tumour. The mitotic count should be included although mitotic figures may be hard to identify in small biopsy samples.

There is little information from clinical trials to allow guidelines to be established for Ki-67 trigger levels. However chemotherapy may be favoured as an earlier treatment option for tumours with Ki-67 of >10%, whereas chemotherapy tends to be used after failure of other therapies in tumours with Ki-67 lower than 5%.[5][6]

The difficulty with using only the Ki-67 labelling index as a guide to determining therapeutic options lies in the difficulty with which the index is assessed histologically. This is due to the following reasons:

  1. The heterogeneity within a NET of Ki-67 labelling. The index should be determined in the areas of greatest staining and an average of several of these fields should be taken.[7]
  2. The difficulty of counting positive staining nuclei and obtaining an accurate labelling index. The ENETs (European Neuroendocrine Tumour Society) committee suggests counting Ki-67 positive nuclei in 2000 cells, which is impractical in a laboratory setting. Ideally, the counting should be conducted on a computer generated image, which is not an easily available facility in a laboratory. A recent review suggested that “eyeballing” the slide is adequate.[7] There is no available data on the degree of inter-observer and intra-observer variability of this assessment, upon which therapeutic options might be determined.

The use of the Ki-67 labelling index alone may not be ideal for determining therapeutic options and the ENETs grading system may be more useful until more information becomes available over time.

Back to top

Standards for classification, staging and grading

Knowledge of the development and biological nature of neuroendocrine tumours has advanced significantly from the days when they were referred to as carcinoids and considered to be a fairly homogenous group of tumours. They are now known to be a heterogenous group of tumours with different behavioural patterns depending on site of origin and cellular characteristics.

WHO 2000 classification of neuroendocrine tumours

The World Health Organization (WHO 2000) classification[8][9] attempted to stratify and separate NETs into behavioural categories according to size, depth of invasion and the presence or absence of angioinvasion. This led to the following tumour categories tabled below.

  1. Well differentiated endocrine tumour
    a) Benign behaviour
    b) Uncertain behaviour
  2. Well differentiated endocrine carcinoma
  3. Poorly differentiated endocrine carcinoma

Recognising that the tumours behave differently according to site of origin, the size and invasion criteria varied between tumours of the stomach, upper and lower GI tract and pancreas.

Other categories included mixed exocrine-endocrine carcinomas and tumour-like lesions.

This classification did not gain uniform acceptance owing to the combination of staging criteria with a grading system, and the difficulties with the acceptance of the category of tumours with uncertain malignant behaviour.

The European Neuroendocrine Society (ENETS) grading system

To address these problems, a histologic grading system based on mitotic rate and Ki-67 labelling index was proposed by the European Neuroendocrine Tumour Society (ENETS), which is tabled below.[3][4]

Grade Mitotic count (10hpf) at x40 magnification Ki-67 index (%)
1 <2 Up to 2%
2 2-20 3-20%
3 >20 >20%


  • 10hpf (high power fields) = 2 mm2, at least 40 fields at 40x magnification evaluated in areas of highest mitotic density.
  • Ki-67 index is a percentage of 2,000 tumour cells in areas of highest nuclear labelling (note comments on Ki-67 labelling above).
  • The grade 2 category identifies and recognises an intermediate group of NETs that shows a greater degree of pleomorphism, mitotic rate and Ki-67 labelling index.

The ENETs classification system incorporates the histological grading system with the tumour stage in recognition of the malignant nature of these tumours and the influence of their biology and stage on the clinical outcome as in other malignant tumours.

WHO 2010 classification of neuroendocrine tumours

This grading system is now incorporated into the new WHO (2010) classification[10] of neuroendocrine tumours with corresponding changes in terminology.

WHO 2000 WHO 2010 Histological features
Well differentiated endocrine tumour NET G1 Well differentiated, mild to moderate nuclear atypia, corresponds to ENETs G1 by Ki67 labelling index and mitotic count
Well differentiated endocrine carcinoma NET G2 Well differentiated, mild to moderate nuclear atypia, corresponds to ENETs G2 by Ki67 labelling index and mitotic count
Poorly differentiated endocrine carcinoma NEC G3

a) Small cell
b) Large or intermediate cell

Poorly differentiated, marked nuclear pleomorphism, necrosis, corresponds to ENETs G3 by Ki67 labelling index and mitotic count
Mixed exocrine-endocrine carcinoma (MEEC) Mixed adenoneuroendocrine carcinoma (MANEC) Malignant tumours with mixed glandular and neuroendocrine characteristics, with at least 30% of one component
Tumour-like lesions (TLL) Hyperplastic and preneoplastic lesions

NET = Neuroendocrine tumour
NEC = Neuroendocrine carcinoma

Back to top

The TNM Staging System for GEPNETS

A new TNM staging system for GEPNETS is included in the 7th edition of the Cancer Staging Manual of the AJCC [1] which should be used for the pathology staging of GEPNETs. The tumour assessment for the stomach, small intestine and large intestine differ and are listed below:

Tumor duodenum.jpg
Tumor colon.jpg

The Node and Metastasis assessment applies to all 3 gastrointestinal regions and is as follows: Lymph nodes.jpg

The anatomic stage and prognostic groups are:
Prognostic groups.jpg

An assessment of the presence or absence of lymphovascular invasion, and clearance of surgical margins should be included in the pathology report.

Pancreatic neuroendocrine tumours

Pancreatic neuroendocrine tumours are graded in the same manner as gastrointestinal tract tumours. For the staging of these tumours, the AJCC[11] adopts the same system as for pancreatic ductal adenocarcinomas (below) which differs from the ENETS stage groupings.

Definition of TNM.png

Note: For pancreatic NETs, the Tis category and explanatory note do not apply.

With the trend towards structured pathology reporting of tumours, consideration should be given to adopting a proforma report for GEPNETs. It may be anticipated that the College of Pathologists of Australasia will introduce such a proforma in the future, but in the meantime, the pathology reporting of a GEPNET should include data in the checklist below.

Checklist for the reporting of GEPNETs:

  • site of tumour
  • multiplicity (given that a checklist applies to one particular pathology specimen episode, multiplicity refers to synchronous multiple tumours identified in the specimen)
  • size
  • differentiation – well or poorly differentiated
  • extent of local invasion/surgical margins
  • presence or absence of lymphovascular space invasion
  • presence or absence of perineural invasion
  • Ki-67 labelling index/mitotic rate
  • lymph node status
  • presence or absence of metastasis
  • surrounding disease e.g. chronic gastritis, IBD
  • somatostatin receptor status where applicable.

The summary of the report should incorporate the WHO 2010 classification of the tumour into NET G1, NET G2, NEC or MANEC, with the TNM tumour stage for the specific site.

Back to top


  1. Srivastava A, Hornick JL. Immunohistochemical staining for CDX-2, PDX-1, NESP-55, and TTF-1 can help distinguish gastrointestinal carcinoid tumors from pancreatic endocrine and pulmonary carcinoid tumors. Am J Surg Pathol 2009 Apr;33(4):626-32 Available from:
  2. Cai YC, Banner B, Glickman J, Odze RD. Cytokeratin 7 and 20 and thyroid transcription factor 1 can help distinguish pulmonary from gastrointestinal carcinoid and pancreatic endocrine tumors. Hum Pathol 2001 Oct;32(10):1087-93 Available from:
  3. 3.0 3.1 Rindi G, Klöppel G, Alhman H, Caplin M, Couvelard A, de Herder WW, et al. TNM staging of foregut (neuro)endocrine tumors: a consensus proposal including a grading system. Virchows Arch 2006 Oct;449(4):395-401 Available from:
  4. 4.0 4.1 Rindi G, Klöppel G, Couvelard A, Komminoth P, Körner M, Lopes JM, et al. TNM staging of midgut and hindgut (neuro) endocrine tumors: a consensus proposal including a grading system. Virchows Arch 2007 Oct;451(4):757-62 Available from:
  5. Jamali M, Chetty R. Predicting prognosis in gastroentero-pancreatic neuroendocrine tumors: an overview and the value of Ki-67 immunostaining. Endocr Pathol 2008;19(4):282-8 Available from:
  6. Vilar E, Salazar R, Pérez-García J, Cortes J, Oberg K, Tabernero J. Chemotherapy and role of the proliferation marker Ki-67 in digestive neuroendocrine tumors. Endocr Relat Cancer 2007 Jun;14(2):221-32 Available from:
  7. 7.0 7.1 .
  8. Solcia E, Kloppel G, Sobin LR. World health organization international histological classification of tumours: histological typing of endocrine tumours: second edition. Berlin: Springer; 2000 [cited 2014 Jun 11] Available from:
  9. Klöppel G, Perren A, Heitz PU. The gastroenteropancreatic neuroendocrine cell system and its tumors: the WHO classification. Ann N Y Acad Sci 2004 Apr;1014:13-27 Available from:
  10. Bosman FT, Carneiro F, Hruban R H, Theise N. WHO classification of tumours of the digestive system, fourth edition. France: IARC; 2010 [cited 2018 Jul 10] Available from:
  11. Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, (editors). AJCC cancer staging manual, 7th edition. France: Springer; 2010 [cited 2014 Jun 12] Available from:

Back to top

Cosa prev button.png
Cosa overview button.png
Cosa next button.png